OK 8a. o o o I?.
Permeability, Porosity, Dispersion-, Diffusion-, and Sorption Characteristics of Chalk Samples from Erslev, Mors, Denmark Lars Carlsen, Walther Batsberg, Bror Skytte Jensen, and
Peter Bo
Risø National Laboratory, DK-4000 Roskilde, Denmark August 1981
RISØ-R-451
PERMEABILITY, POROSITY, DISPERSION1-, DIFFUSION-, AND SORPTION
CHARACTERISTICS OF CHALK SAMPLES FROM ERSLEV, MORS, DENMARK
Lars Carlsen, Walther Batsberg, Bror Skytte Jensen, and Peter Bo
Chemistry Department
Abstract. A series of chalk samples from the cretaceous forma
tion overlying the Erslev «;alt dome have been studied in order
to establish permeabilities, porosities, dispersion-, diffusion-,
and sorption characteristics of the chalk, Predominantely the
investigations have been carried out by application of a liquid
chromatographic technique.
The chalk was found to be porous (c z. 0.4), however, of rather _7
low permeability (k z. 10 cm/sec). It was found that the material exhibits a retarding effect on the migration of cationic
(continue on next page)
August 1981
Risø National Laboratory, DK 4000 Roskilde, Denmark
species as Cs , Sr , Co , and Eu , whereas anionic species as
Cl~ and TcoT move with the water front.
The geochemical implications are discussed.
INIS descriptors; ADSORPTION; CALCIUM CARBONATES; CESIUM 134;
CESIUM COMPOUNDS; CHALK; CHLORINE 36; CHLORINE COMPOUNDS; COBALT
60; COBALT COMPOUNDS; DIFFUSION; EUROPIUM 154; EUROPIUM COM
POUNDS; EXPERIMENTAL DATA; FLOW RATE; GEOCHEMISTRY; ION EXCHANGE;
LIQUID COLUMN CHROMATOGRAPHY; MIGRATION LENGTH; PERMEABILITY;
PERTECHNETATES; POROSITY; POROUS MATERIALS; RADIOISOTOPES; RADIO
NUCLIDE; MIGRATION; STRONTIUM 85; STRONTIUM COMPOUNDS; TECHNETIUM
99; WATER.
UDC 552.54
ISBN 87-550-0785-6
ISSN 0106-2840
Risø Repro 1981
CONTENTS
Page
1. INTRODUCTION 5
2. BASIC PRINCIPLES OF LIQUID CHROMATOGRAPHY 6
3. EXPERIMENTAL 10
3.1. Preparation of Chalk columns 10
3.2. Liquid chromatography equipment 11
3.3. Density determinations 12
3.4. Retention data by batch-type experiments 13
I. RESULTS 14
4.1. Permeability 14
4.2. Porosity 17
4.3. Dispersion 20
4.4. Diffusion coefficients 21
4.5. Sorption phenomena 25
5. GEOCHEMICAL IMPLICATIONS 38
5.1. Redox and pH conditions in the Erslev chalk
formation 38
5.2. Geochemistry of radionuclides 42
6. CONCLUSIONS 44
ACKNOWLEDGMENTS 46
REFERENCES 46
- 5 -
1. INTRODUCTION
In connection with the possible disposal of nuclear waste in the
salt dome located at Erslev, Mors, Denmark, a physico-chemical
characterization of the chalk formation overlying the salt dome
has been undertaken.
The chalk samples investigated in the present study originates
from the Frslev wells ERSLEV IS, ERSLEV 2S, ERSLEV 3S, and ERSLEV
4S, respectively. The samples were received from the Danish Geo
logical Survey (DGU). A detailed geological characterization of
the chalk formation has been performed by DGU, the results being *)
described elsewhere.
In the present study a series of physico-chemical measurements
on the chalk samples (in total 21 samples have been investigated)
including determinations of permeabilities, porosities, disper
sion, and sorption characteristics are reported, supplementary
with a discussion of the results obtained. In addition a discus
sion of the posible geochemical implications of the derived data
is given.
In the major part of the experiments a modified liquid chroma
tographic technique has been applied to determine permeability,
porosity, dispersion, and retention data for the chalk samples.
Additionally, the porosities have also been calculated on the
basis of density measurenents, and the sorption phenomena studied
by batch type experiments.
*) The geological characterization of the Erslev chalk formation
has been reported by the Danish Geological Survey as a con
tribution to the ELSAM/ELKRAPT management project, phase 2.
f>
2. BASIC PRINCIPLES OP LIQUID CHROMATOGRAPHY (LC)
Liquid chromatography is a column technique used for the separ
ation of different organic as well as inorganic components in a
solution. The separation mechanism depends on the nature of the
physical/chemical interaction between the solute and the column
packing Material. For a given column material, e.£. chalk, the
separation mechanism will he controlled by adsorption as well as
ion-exchange processes. As eluent flow through the column, dif
ferent sorites can selectively be retarded, but still eluted. A
chromatogram is obtained when the outlet fro* the colusm is moni
tored by a e.g_. refractive index detector or a radioactivity
monitor. A typical experimental set up for & liquid chromato-
grafic separation is visualized in Pig. 1.
ELUENT RESERVOIR a
INJECTION PORT ¥ PUMP
COLUMN
RECORDER I
DETECTOR
r WASTE
Fig. 1. Experimental liquid chromatography a*t up.
The pump delivers a solvent (eluent) flow at a preset constant
flow rate. A sample of known volume is injected into the column
via the injection port, and the actual composition of the eluate
from the column is examined by the detector.
- 7 -
Based on the appearance of the chromatogram a variety of physico-
chemical characteristics for the column can be calculated.
The PERMEABILITY, K, for a given column is determined by the vol
ume flow rate, Q (cm /sec), the length of the column, 1 (m), the 2
cross sectional area of the column, A (cm ), and the pressure
drop over the column, h (metre of water) (Eq. 2.1) (Dahl, 1979).
K * Q-l/A-h (cm/sec) (2.1)
The volume porosity, e, of the column can be determined by injec
tion of a solute, which is not retarded by the column packing
material (Eq. 2.2)
c = V 0/A1 (2.2)
where V i s the so-ca l led dead volume (cf. Fig. 2 ) .
UNRETAROED SOLUTE
INJECTION A RETARDED OF SAMPLE fil SOLUTE
VxO V0 VR VOLUME t=0 t 0 tR TIME
Pig. 2. Schematic chroaatograa defining the elution voluae (V ) and elution tiae {t l for the unretarded solute and the retention o voluae (VR) and retention tiae (tR) for a retarded solute.
The shape of the peak corresponding to an unretarded solute characterizes the flow dispersion in the column, whereas the elution volume (or elution time) and shape for a retard«i peak (cf. Fig. 2) characterizes the sorption phenomena, related to the solute under investigation,to the column packing material.
The FLOW DISPERSION, o, i s determined from the width of the chroma togr am, w, measured at the base l ine . For a chromatogram, exhi-
- t -
biting a Gausian distribution with a standard deviation, o, the
o-w dependence is given by Eg. 2.3 (Yau, Kirkland and Bly, 1979).
w * 4-o (2.3)
The RETENTION FACTOR is determined according to Eq. 2.4 (cf.
Pig. 2)
*f * V*« * V*I (2.4)
The sorption ability of a given colusn packing material is in
general expressed by the distribution coefficients. Kp, for the
solute N (Eq. 2.5)
where c„ and c„ are the concentrations of M on the solid phase
and in solution, respectively. The K_ values are related to the
retention factors by the expression given i Eq. 2.6
Rf - (1 • (1-cJ-p-Kjj/e)"1 (2.6)
Nhere e is the above Mentioned voluae porosity and p is the bulk
density of the solid phase. The retention factors reflect the
novenent of N relative to the solvent front.
In the major part of the experiments reported here the LC-tech-
nique are applied to so-called homogeneous columns consisting of
undisturbed chalk samples (cf. experimental section), and the
solutes are selected so that both the sorption properties of
relevant radionuclides as well as some hydraulic parameters of
consolidated chalk from Erslev can be determined.
However, although studies of retention phenomena in principle
can be performed on homogeneous columns, it shall be emphasized
that retention times, t_, higher than 1 0 - 1 5 times the dead
time, tQ, of the column (.i.e. retention factors less than ca.
0.07) afford unsatisfactory broad peaks, which at the limit may
- 9 -
escape detection. For the study of sorption phenomena, resulting
in retention factors less than ca. 0.07 so-called heterogeneous
columns say advantageously be used. These columns consist of a
known amount of crushed Material, which sorption characteristics
is to be investigated« nixed with "hydrophilic" polystyrene
spheres with a nean particle diameter less than 160 in, the lat
ter exhibiting no retarding effect towards the radionuclides
studied in the present study. Since the retarding effect of these
"diluted" columns apparently is considerable less than that of
the corresponding honegeneous columns, these types of experiments
enable us to study the retention of radionuclides with such
saaller retention factors than do those with the homogeneous
colussis.
It is importart to net.2 that the study of retention phenoaena in
the heterogeneous colusms give rise to determination of apparent
retention factors, R°rs, according to Eq. 2.4. It shall, however,
be emphasized that the true retention factors, R-, corresponding
to the retention in a column consisting of undiluted, undisturbed
sorbing material, are not directly obtainable, since the dilution
with non-sorbing material affords a decrease in the apparent dis
tribution coefficients, K ? , proportional to the dilution (Eq.
2.7)
Kj*» - V « " 1 (2.7)
where K° b s is th« distribution coefficient in the diluted ma
terial and a is the dilution factor (a£l; in the present study
dilution factors between 10 and 20 have been applied).
By analogy with Eq. 2.6 the connection between R? and l£ is
given by Eq. 2.8
R|b" - (1 • d-c').p'.KjbVe')-1 (2.8)
which may be rewritten into the following expression for x£b*
(Eq. 2.9)
Kjb» - ((1-R|bs)/R|b,)-eV(1-c').p' (2.9)
- 10 -
c" and ø* are prosity and bulk density of the heterogeneous
column. Combining 2.7 and 2.9 we obtain the expression (Eq. 2.10)
for the true distribution coefficient
Kp » a-tcVpMI-c'))«!-«!6*)/«!1** (2.10)
The true retention factor can be calculated according to Eq. 2.6.
It should be noted that Rf factors determined dynamically be use
of heterogeneous coluans correspond to retention factors on pure
crashed material* and may differ from the true values in the con
solidated material (cf. the discussion on batch type vs. col«
type experiments in section 4.5).
3. EXPERIMENTAL
The chalk samples were received from the Danish Geological Survey
as cylindrical specimens with a diameter approximately 10 cm and
a length of approximately 12 cm. Some of the samples were waxed
and wrapped in a heavy polyethylene foil, others only packed in
polyethylene foil.
3.1. Preparation of chalk columns
Immediately after unwrapping, the cylindrical specimen was cut
twice using a band saw. Two half cylinders and a rectangular
plate with nearly the same dimensions as the length and the di
ameter of the cylinder were obtained. A small part of the plate
was used for the chemical analyses and water determinations (Sol
gård and Skytte Jensen, 1981). From the remaining part of the
plate, rods with the dimensions 120 x 12 x 12 mm were cut either
parallel or perpendicular to the cylinder axis. The rod was
mounted in a lathe and while rotating polished to cylindrical
rods with a diameter of 10 mm. Finally, they were cut to the
length of 100 mm, 50 mm, or 10 mm.
- 11 -
Before use in a coluan experiment the cylindreal chalk rod
provided with end fittines of polyethylene, porous Metal and glas
fiber filters and encapsulated in a heat shrinkable tubing with
a dianeter of 12.5 an. after heat shrinking of the tubing, the
coluun wes further encapsulated in another heat shrinkable
tubiig. after beat shrinking the total outer dianeter of the
cciumn was approxi.jately 12.5 na fcf. pig. 3}.
"lflr"***f .
CO A 0 C
>B1 Nvle.«
Chalk ral—u as t»* •e mniiiiiwi, sc r»ir-
Occasionally it was inpossible to cut out a rod with a total
length of "00 m , in which cases the colunms were nade of nore
than one piece. For samples with very low pemeabilities or for
the determinations of retention factors, the total length of the
sanples was reduced to 50 mm or 10 mm. In these cases cylindri
cal teflon rods with a bore of 1.5 m were used as spacers in
order to retain the total length of the colunm.
3.2. Liquid chromatography equipment
For the determination of permeability, pore volume and disper
sion a high pressure pump, model FK-30, KMAUEft KG, Berlin, BUD,
an injection port, model 7120, Kheodyne, California, USA, pro
vided with a 20 tiL loop, a compression module, BOH 100, waters
Ass., Mass., USA, retrofitted with an accurate manometer, and a
refractive index J«tector, model K-401, Maters Ass., Mass., USA,
were used (cf. Fig. •}. For the retention measurements the KMAUER
pump was substituted with a Kontron pump, model LC-410, and the
detector was a radioactivity monitor LB-503, Or. Berthold, Wild-
bad, BRD. The injection port was for the retention measurements
provided with a 50 »L loop. In both cases a BD-9 recorder, Kipp-
- 12 -
Zonen, 'as used and the flow rate was determined by collecting
the eluate in a graduated cylinder for a given time.
The experiments were performed in the following way: The chalk
column was inserted in the RCM-100 compression module (cf. Fig.
4), and a hydrostatic pressure on the column was generated by
activating the different levers. Redistilled water or salt sol
utions was pumped througt the column. The flow rate was adjusted
so that the hydrostatic pressure in the compression module was
higher than the pressure drop over the column. Before injection
of the solutes the column was equilibrated untill a stable base
line of the detctor signal was obtained.
Fig. 4. RCM-100 Compression Nodule. Lowering the levers applies pressure to the column (courtesy of Waters Ass., Denmark).
3.3. Density determinations
The densities of the chalk samples were determined by weighing
the samples in air and immersed in redistilled water. Before
weighing the samples were equilibrated overnight in redistilled
water.
- 13 -
3.4. Retention data by batch-type experiments
Dry crushed samples (ca. 250 mg) were equilibrated in centrifuge
tubes for 1 hour with 26 mL of a salt solution (salt solution A:
25 mL 1M NaCl + 1 mL radionuclide solution; salt solution B: 25
mL 5M NaCl + 1 mL radionuclide solution. The radionuclide sol-3+ 2+ 2+ + ution contains Eu , Sr , Co , and Cs , all in concentrations
of ca. 2.5«10~ eq/L, and trace amounts of Eu *, Sr +,
Co + , and Cs , corresponding to ca. 0.2 uCi/mL).
After centrifugation (J h) 10 mL of the supernatants were trans
ferred into counting bottles; the remaining supernatant was re
moved by decantation. The moist chalk samples were weighed in
order to correct for the amount of solution which cannot be re
moved in this way. The moist chalk samples were then dissolved
in 10 mL 2M hydrochloric acid, the resulting solution being
transferred to counting bottles.
The samples (supernatant and chalk) were then counted on ND-100
multichannel analyzer using Ge(Li) detector ( Eu: 122.keV, Sr:
513 keV, 134Cs: 604 keV, 60Co: 1173 keV). Distribution coef
ficients were calculated according to Eq. 3.1
KD - 5 -A f ^ a,L/« (3-1) Cfd sol
where W„ ,, and W„ _ ara the wights of the dry and moist chalk c f o c fin
samples, respectively, and AcnalJc and A g Q l are the radioactivity
(counts/1000 sec.) for the dissolved chalk samples and 10 mL of
the supernatant, respectively.
The calculations of K_ values are carried out separately for each
of the four nuclides.
All determinations were carried out in duplicate.
- 14 -
4. RESULTS
In the following section the results for permeability, porosity,
dispersion, diffusion, and retention will be given separately.
4.1. Permeability
The permeability is calculated from Eq. 2.1. Correct values of
the permeability is obtained if the presence of voids or chan
nels at the interface between the chalk rod and the polyethylene
tubing can be excluded. This will be the case when the hydro
static pressure in the compression module is higher than the
pressure drop over the column. In Table 1 the permeabilities as
function of flow rate and hydrostatic pressure for a given chalk
column is given.
Table 1. Permeability (K), volume porosity (e), and disper
sion (a) as function of hydrostatic pressure (P ) and press
ure drop (P.) over 100 mm Erslev 1s-5 column.
P PA flow rate permeability volume porosity dispersion
bar bar mL/min cm/sec % n
40 42 46 56 66
84 84
133 145
21 31 42 52 70
33 78
66 135
The results (Table 1) show that the permeability is nearly inde
pendent of the pressure drop over the column over a wide range
of flow rates. If very high hydrostatic pressures are applied,
0.101 0.167 0.235 0.293 0.360
0.170 0.360
0.167 0.360
1.02-10 1.14-10 1.19-10 1.20-10 1.09-10
1.09-10 1.00-10
0.55-10 0.57-10
40.5 40.2 41.1 40.1 33.7
41.4 41.3
38.3 38.0
0.052 0.058 0.056 0.061 0.065
0.058 0.059
0.053 0.051
- 15 -
the permeability appears to decrease slightly. A similar trend
is observed in the values for the pore volume and the dispersion.
The permeabilities as determined for all 21 chalk samples avail
able are visualized in Fig. 5 and 6 as function of depth below
ground level. All data were obtained at flow rates where P > P..
10 "
o
^10" 7
E u
10 -
10 k-9
-
-
1
' 1
t
I
' 1 1
Erslev 2 S /
t
• i i
i ' i
/Erslev IS
-
i . i
600 500 400 m
300 200
Fig. 5. Permeabilities as function of aepth for chalk samples from Erslev 1s and 2s.
All samples except the two marked t were cut perpendicular to the
well direction. The permeability of each of the two samples
(Erslev 1s-10 and Erslev 1s-12), cut parallel to the well direc
tion, is comparable to that for a sample from Erslev 1s-11.
Supplementary, a determination of the possible change in per
meability as function of time for a column exhibiting a well de
fined crack was performed. The first 30 mm of the column in the
direction of the flow was a normal consolidated cylidrical rod,
whereas the nex 70 mm was made of two half cylinders. As the
- 16 -
-6
10
o tf) -7
" 107 E o
IQ'8
600 500 400 300 200 m
Pig. 6. Permeabilities a« function of depth for chalk samples
from Erslev 3s and 4s.
eluent (redistilled water partially saturated with calcium car
bonate) was passed through the first 30 mm, a further saturation
of the eluent with CaCO? was assumed to occur. The hydrostatic
pressure on the column was approximately 100 bar through out the
experiment. The permeability was determined at different times,
and the results are given in Table 2.
At the end of the experiment (160 hours) approximately 600 mL
eluent have been passed through the column, corresponding to ca.
150 times the pore volume in the column. Only minor variations
in the permeability is observed. Especially the surprisingly low
initial permeability should be noted.
1 1 I ' i i 1 • T
17 -
Table 2. Permeability of a chalk column initially containing
a well defined, macroscopic crack. (Erslev 4s-3, flow rate:
oa. 0.06 mL/min., P ca. 100 bar, P. ca. 30 bar)
tine (after start of exp.) permeability
4.2. Porosity
hours en/sec.
0 30 45 117 140 160
5.7-10 4.2-10 4.7-10 4.5-10 5.3*10 4.7*10
-7 -7 -7 -7 -7 -7
The pore volumes of the chalk columns was determined from a
chromatograra obtained after injection of a 20 yL sample contain
ing 30% D20 (cf. Pig. 7). It is assumed that D20 is not retarded
in the chalk columns, equilibrated with H_0 before the injection
of the D20/H20 sample. However, if the permeability of the column
is less than ca. 10~ cm/sec. the flow rate through the column
Por« vol:38.29 Oisp.:0.O52 xV
i
2.906 ml
3.156 ml
A
f0.636'ml
Fig« 7. Determination of pore volua* and disportion froa the ebroaatooraa of a DjO solution (10 ca Erslev le-5 eolaan (flow rata 0.1C7 al./ain., chart speed 0.5 ca/nin.)).
- 1 3 -
is so low that lack of stability cf the refractive index detector
for the long elution times needed seems to be prohibitive for a
good determination of the pore volume. Hence, the elution times
can be reduced by introduction of a shorter column (50 mm or 10
mm), however, excessive external volumes (e.g. dead volume in
the spacers, cf. Experimental section) have to be taken into ac
count .
For columns with low permeabilities the pore volume can advan
tageously be determined by injection of a radioactive sample,
e-2« Cl~, which will not be retarded on the column, and sub
sequent detection with a radioactivity monitor.
50
40
30
20
50
30
20 600 500 400 300 200 100
m
Tig. 6. Porosities as function of depth for chalk samples from Erslev Is and 2s.
V From chromatogram • Assuming H20 in the pores o NaCI solution in the pores (
a Water in the pores
V From chromatogram i Assuming H20 in the pores o NaCI solution in the pores a Water in the pores
Erslev 2s
- 19 -
Porosities were additionally determined from density measurements
on chalk samples with the original salt solution in the pores or
with redistilled water in the pores (.i.e. samples primarily used
for permeability measurements were applied here). The porosities
were calculated assuming the density of non-porous CaCO, to be 3 J
2.7 g/cm , and a density of the original salt solution of 1.00
g/cm . Complementary, the actual density of the salt solution
originally present in the pores, the molarity being established
in a supplementary stud/ (Solgård and Skytte Jensen, 1981), was
used. In Fig. 8 and 9 tne results are summarized.
50 h
* 40
aP
30
20
50
40
30
20
V From chromatogram Assuming H20 in the pores
o NaCi solution in the pores D Water in the pores
Erslev 3s
_ V From chromatogram • Assuming H20 in the pores o NaCI solution in the i a Water in the
600 500 400 300 m
200 100
Fig. 9. Porosities as function of depth for chalk samples from
Erslev 3s and 4s.
- 20 -
It is noted that a good agreement between the "NaCl" (o) and the
"H.O" (D) results is obtained. The porosities determined from
the chromatogram are in general somewhat lower.
4.3. Dispersion
In Table 1 the dispersion as function of flow rate for a 100 mm
Erslev 1s-5 column is given, the standard deviation of the chrom-
atogran, o, being calculated in units of the pore volume, V . (Fig. 7)
If the dispersion observed is due to diffusion only, the Einstein-
Smoluchowski relation (Eq. 4.1) is valid (Atkins, 1978).
0-L « /2D^t (4.1)
where L is the length of the column, D is the diffusion coef
ficient of D20 in HjO, and t is the retention time. Inserting
the actual values, L < 10 cm, o = 0.052, and t = 31.5 min into
Eq. 4.1 the diffusion coefficient D is calculated to be 7.10 2
cm /sec, which is somewhat higher than the gennerally accepted —5 2
values for self diffusion in water (10 cm /sec). However, cor
rections due to dispersion phenomena in the injection system, the
connecting tubings, and the detector systems have to be taken
into account. Hence, the corrected value is calculated to be -5 2
6.2*10 cm /sec, on which background we are forced to conclude
that other processes than diffusion contributed to the disper
sion of D,0 in the chalk column at the flow rates examined
(>0.03 mL/min). An identical result is obtained after injection
of the non-retarded radionuclide Cl~ followed by monitoring
the eluate with a radioactivity detector. It should in this con
nection be emphasised that the lower limit of flow rate by these
experiments (0.03 mL/min) most surely far exceed the actual flow
rate in the chalk formation, and that a further lowering of the
flow rate through the chalk columns is assumed to afford a sub
sequent decrease in the dispersion coefficient. Due to the very
low permeabilities the effective diffusion coefficient of el
ements in the pore water is believed to be up to several orders of magnitude below the diffusion coefficient in pure water (10
2 cm /sec).
- 21 -
4.4. Diffusion coefficients
In the present study determination of the diffusion coefficients
for sodium ions have been deterained for two representative chalk
samples, Erslev 3s-6 (-325 m) and Erslev 4s-7 (-552 m ), respec
tively.
4.4.1. Diffusion in porous media
In porous media the apparent diffusion coefficient« D, measured
relative to the open ends of a pore, is less than the intrinsic
diffusion coefficient in the pore fluid by a factor equal to the
square of the tortuosity of the pore, since the actual path is
increased proportionally to the tortuosity, whereas the concen
tration gradient along the pore is reduced proportionally to the
tortuosity. According to the theory of Mackie and Meares the ap
parent diffusion coefficient, D, is given by the empirical Eq.
4.2, where D„„ refers to the diffusion coefficient in pure water aq
(Meares, 1968).
D = Daq(e/(2-e))2 (4.2)
e is the porosity of the media.
However, it should be emphasized that the Eq. 4.2 applies to
homogeneous media and only when the pore diameter is large com
pared to the diameter of the diffusion molecule.
A distinct decreasing effect of the apparent diffusion coef
ficient with decreasing pore size has been observed (Beck and
Schultz, 1970), leading to a revised version of the above equa
tion for D, given by Eq. 4.3.
D « Daq.Q-(e/(2-e))2 (4.3)
where Q is a constant less than 1. The actual magnitude of Q is
solely dependent of the nature of the porous media and the dif
fusing material. In the present case it will accordingly be ex
pected that for a given chalk sample, the factor Q will decrease
- 22 -
with increasing size of the diffusing ions. Analogously it will
be expected that Q will decrease for a given ion, if the pore
size of the chalk samples investigated decreases, .i.e. the per
meability of the samples decreases.
4.4.2. Diffusion coefficients in pure water
The diffusion coefficients for ions, i, in pure water, D , can aq
be det rmined by application of the limiting ionic conductances,
X± (Q~ cm2), in water (Eq. 4.4) (Harned and Owen, 1963)
Daq " *i*T/!*i!
2F2 (cm2s_1) (4.4)
where R is the gas constant, T is the absolut temperature, z. is
the ionic charge, and F is the Faraday.
According to Eq. 4.4 the diffusion coefficients of sodium ions
in water at room temperature can be calculated to be 1.33*10
cm s .
4.4.3. Apparent diffusion coefficients in chalk samples
For diffusion coefficient measurements the column in the liquid
chromatographic system, described above, was substituted by a
"diffusion unit" (Fig. 10) consisting of a 6 mL reservoir (fill
ed with eluent) through which the eluent can be passed with a
preset flow rate. At the top of the diffusion unit the column,
still encapsulated in the shinkable polyethylene tubing is placed
as visualized in Fig. 10. However, the end fittings in the one
end were removed, to ensure a free chalk surface in contact with
the eluent in the reservoir, ^.e. the contact area equals the
cross sectional area of the chalk column. The upper end of the
column is closed in order to prevent a possible flow up through
the column. The eluent in the reservoir is stirred to ensure a
uniform concentration within the total volume and avoid concen
tration polarization at the surface of the chalk. The variations
in concentration in the reservoir as a result of a diffusive
leaching from the column is followed by the detector by passing
eluent slowly through the diffusion unit.
- 23 -
Fig. 10. Diffusion unit.
In the present case the chalk columns were, prior to the dif
fusion experiments saturated with a 1M Nacl solution containing 22 • Na , using the above described liquid chromatographic system.
The diffusion unit was eluted with pure 1M NaCI, and the vari-22 •
ations in Na in the reservoir were followed by the radioac-22 •
tivity monitor. No sorption of Na on the chalk was observed,
possibly due to the high sodium concentration in the liquid
phase.
- 24 -
Th« concentration variations obtained in this way were then simu
lated by application of the computer program DIFMIG (Bo and
Carlsen, 1981a), which is designed to handle diffusive rigration
of radionuclides through column systems. In order to obtain sat
isfactory simulations it was necessesary to include diffusion co
efficients in the near surface region of the column (up to 1 am)
which are somewhat higher than the actual diffusion coefficient
used for inner part of the column« presumably due to surface im
perfections .
The apparent diffusion coefficients, D, for Tta* have been de
termined for two representative chalk samples from the Erslev
formation. The results are summarized in Table 3, the data for
the porosities, c and permeability, K, being adopted from above.
Table 3. Diffusion data for chalk samples from the Erslev
formation.
sample e K (cm s"1) D (cm2 s'1)
Erslev 3s-6 0.4 ;.6-10~7 5.0«10~8
Erslev 4s-7 0.3 1.7-10"8 1.7-10"8
Fig. 11 shows the experimentally obtained variation in 22-a+ con
centration (?) and the corresponding DIFMIG simulation.
For the two chalk samples the diffusion coefficients in the near —6 2 —1
surface region ware: Erslev 3s-6: 3.4*10 cm s (0-0.4 ma) and
2.2-10"6 or^s"1 (0.4-0.8 am). Erslev 4s-6: 1.2-10"6 crn̂ s"1
(0-0.4 ma) and 1.0-10"6 ca2*"1 (0.4-0.8 am). As seen these coef
ficients vary, as expected, with the column used and are not ex
pected to be reproducible, owing to differences in surfaces as a
result of th« handling of th« individual samples. They should,
however, always be less than the intrinsic diffusion coefficients 22 • of Na i.i pur« water.
- 25 -
r u . l i . Diffusive itachlaf cf 2XSa* fraa i i t a Erslev Is-« colnm. h f i r I — f l iy åmimwutwmå c n c w t n t K M «f elw iMchi lexerasss« la CSM nm flvea sy v, the w l M ran« tmtmq ttm
tlcal siaslatlsa *T 0Z7MZC.
Pros) the above date (Table 3) i t i s possible to estimate the factor Q (cf. Bq. 4 .3) . Por the Erslev 3s-« and Erslev 4s-7 the factor 0 were determined to be 0.0(2 and 0.042, respectively. The decrease in Q as a result of decrease in permeability should be noted.
4 .5. Sorption phenomena
The possible sorption of a series of radionuclides in trace amounts onto the chalk has been studied by two distinct different methods: a) batch-type experiments, where crushed chalk samples are equilibrated with a sodium chloride solution, containing the radionuclides, by shaking or b) column-type experiments, where a
i l l sample of the radionuclide solution i s injected into the
- 26 -
LC-system (cf. Fig. 1), the column being continuously eluted with
the sodium chloride solution (saturated with calcium carbonate)
(cf. experimental section).
The radionuclides used for the present sorption studies include
four cations (134Cs+, 60Co2+, 85Sr2+, and 154Eu3+) and two a 36 - 99 -
anionic species ( CI and Tc04)'.
Apart from the importance of cesium, cobalt, strontium, and tech
netium as actual components in nuclear waste, strontium rep
resents group II elements such as barium and radium, whereas co
balt is believed to be a proper representative for other divalent
metal ions- e.g_. iron and nickel. Europium is included in the
study as a reasonable representative for trivalent actinides as
americium (III) and plutonium (III) , the latter most surely
existing as such under the actual conditions within the Erslev
chalk formation (cf. section 5) (Skytte Jensen, 1980). It is, on
the other hand, not inteligible whether cesium may be regarded
as a representative for other monovalent metal ions, such as
rubidium, or whether the sorption phenomena observed in the
cecium case should be connected specifically to the latter, since
rather distinct features connected to cesium sorption have been
reported previously (Grim, 1968).
It is well established that the stable oxidation state of tech
netium in aerated solutions is +7, .i.e. technetium appears under
such conditions as negatively charged pertechnetate ions (TcoT)
(Skytte Jensen, 1980). However, at the actual oxidation poten
tial in the Erslev chalk formation (Eh approx. -0.3; cf. section 2+ 5) technetium may appear in the oxidation state +4, as TcO- .
2+ Furthermore, at pH around 9.0 TcO- will most efficiently precipitate as TcO(OH)2, which, of course, will be in equilibrium with TcO(OH)2 in solution (Skytte Jensen, 1980). Laboratory experiments with these reduced technetium state have not been included.
- 27 -
85Sr2+,
4.5.1. Batch-type experiments 134 + By the batch-type experiments the sorption of Cs
and Eu has been studied. These experiments were carried out
for all 21 chalk samples available. Two series of experiments
were performed, corresponding to bulk sodium chloride concen
trations of IM and 5M for the equilibrating solution, respect
ively. Distribution coefficients were determined as described in
the experimental section (section 3) and the corresponding re
tention factors were calculated by Eq. 2.6 adopting the bulk
CaCO, density as 2.7 g/cm and the porosities reported in sec-
6 0 C o 2 + ,
10l
10 -1
10"2-
10 r3
10"
10 -5
_ Erslev 1S (equil:1M NaCl)
(V 15*Eu3*. o 85Sr2': A 13ACs-: D 60Co2+)
100 200 300 400 500
DEPTH (m) 600
Fig. 12. Retention factoxs a« function of depth (Er»l«v 1g/1M NaCl).
- 28 -
tion 4.2. In Figs. 1 2 - 1 9 the results are visualized, the cal
culated retention factors being depicted as function of depth
below ground level.
10(
10-1-
10~2-
10 i-3
10 -4
10 -5
Erslev 1 S (equil:5M NaCl) (V 15AEu3*; o
85Sr2} A mCs* ; o ^Co2*)
1 100 200 300 400 500 600
DEPTH (m)
Fig. 13. Retention factors a« function of depth (Srsitv 1s/5M NaCl),
In accordance with previously obtained experience (Bo and Carl-
sen, 1981b) values obtained by batch-type experiments are rather
scattered. However, a clear-cut pattern is unambiguously recog
nized concerning the mutual location of the single radionuclides
in the "Rf - depth" diagrammes. In addition a slight tendency to
decreasing F^ values with increasing depth is observed.
- 29 -
10(
10-
io-2 -
10 ,-3
10*4-
10 -5
Erslev2S (equil: 1M NaCl) (y 154Eu3*;o
85Sr2;. A 134Cs*; a ^Co2*)
J_ ± ± -L 100 200 300 400
DEPTH (m) 500 600
Flg. 14. Retention factors as function of depth (Erslav 2s/1M NaCl).
Most striking, however, is the apparent equality of the four
wells. No dramatic changes in the retention factors are observed
comparing the single samples. Accordingly, mean values for the
retention factors for the single radionucledes (including all 21
samples) and standard deviations are calculated (Table 4). The
corresponding mean X. values (adopting e * 0.4 and p * 2.7) are
given in Table 5.
- 30 -
10(
10 -1
10 -2
10 -.3
1Q-
10 -5
_ Erslev2S(equil:5M NaCI) (V 154Eu3*.o 85Sr2*. A 13ACs*. D 60Co2*)
100 200 300 400 500 600 DEPTH (m)
ytg. 15. Retention factors *M function of depth (Erslev 2s/5H NaCI).
Table 4. Mean retention factors and standard deviations.
[NaCI] 1M 5M
R£(Eu)
R f(Co)
R f (Sr)
R*(Cs)
(5 .5 t
( 7 . 8 ±
( 8 . 9 i
(1 .4 ±
3 . 8 ) - 1 0
4 . 4 ) ' 1 0
3 .1) '10
1 .3 ) -10
-5
-3
-2
-1
(5.2 ± 4.4)«10
(9.0 t 4.1)»10
(9.8 ± 3.6)'10
(2.1 ± 1.3)-10
-5
-3
-2
-1
- 31 -
10<
10-V
i o - 2 -
10"
10"
10"
i r
Erslev3S(e^uil:1M NaCl) (V 15*Eu3*;o
85Sr2*; A mCs' ; D ̂ Co2*)
_L 0 100 200 300 400 500 600
DEPTH (m)
Fig. 16. Retention factors as function of depth (Erslev 3s/1H NaCl).
Table 5. Mean K values.
[NaCl] 1M 5M
KD(Eu)
KD(Co)
KD (Sr)
KD(CS)
4 . 5 - 1 0
3 .1 -10
2 .5
1.5
4.8-10-
2.7-10
2.3
0.9
1
- 32 -
10(
10 -1
T0"2h
10"
10 -4
10-
_ Erslev3S(eguil:5M NaCl) tø 154Eu3*. o 85Sr2*: A
13ACs*. D ^Co2*)
100 200 300 400
DEPTH (m) 600
Fi?. 17. Retention factor« as function of depth (Erslev 3s/5M
SaCl).
A slight tendency to increasing R. factors (decreasing K_ values)
by increasing bulk sodium chloride concentration is noted. The
effects of the bulk salt concentration is further discussed be
low in connection with the column experiments.
- 33 -
10<
10 -1
i o - 2 -
10 r3
10"4-
10 -5
ErslevAS(equil:1M NaCI) (V 15AEu3*. o 85Sr2'. A mCs*. a ^Co2*)
i . 100 200 300 400
DEPTH (m) 500 600
Pig. 18. Retention factors as function of depth (Erslev 4s/1M NaCI).
- 34 -
10(
10-1
KT 2 -
10 r 3
io-4-
10"
_ Erslev4S(equil:5M NaCI) (V 15AEu3-; o »Hr2*; * mCs* ; a »»Co2*)
X 100 200 300 400
DEPTH (m) 500 600
Pi*. T*. fcttantion factors as function of depth (Erslev 4s/SH •aCl).
4.5.2. Column-type experiments
The column-type experiments were carried out by application of
the LC-technique. SO uL samples of ca. 10* eq/L solutions of
NaCl, CsCl, SrCl2, and CoCl2 containing trace amounts of Cl", 134Cs*, 85Sr2*, and 60Co2*, respectively, and of K99Tc04 were
injected and the radioactivity eluted from the columns was de
tected by a radioactivity monitor (cf. experimental section).
The columns were eluted with 0.1, 0.5, 1.0, 2.0, and 4.0 molar
sodium chloride solutions, primarily saturated with calcium car
bonate. Since Cl" will not be retarded, the appearance of the
- 35 -
Cl~ signal visualizes the passage of the solvent, front (corre
sponding to t ).
134 • 85 2+ 99 — In the present study the retention of Cs , Sr , and TcO.
has been investigated on homogeneous columns, whereas the het
erogeneous approach has been applied to Co retention.
The determined retention factors for Cs* and Sr * for four
representative core samples are given in Table 6. No retarding 99 -effect of the chalk samples on TcO. was observed.
Table 6. Column-type experiment derived retention factors
for Cs* and Sr ( 10 am homogeneous chalk columns)
ls-5 4s-3 2s-9 4s-6
,,,„, 134_+ 85_ 2* 134„ • 85_ 2+ 134„ • 85-, 2* 134_ • 85_2+ [MaCl] Cs Sr Cs Sr Cs Sr Cs Sr
0.5 1.0 2 .0 4 .0
0.13 0.23 0.40 0.35
0.11 0.24 0.31 0.22
0.12 0.26 G.38 0.51
0.15 0.21 0.28 0.28
0.09 0.12 0.28 0.34
0.11 0.08 0.26 O.18
0.05 0.10 0.17 0.29
0.12 0.14 0.18 0.18
By comparing the values in Table 6 with the average Rf factors
derived from the batch-type experiments (Table 4) it is clear
that the column-type experiments give somewhat higher values than
do the batch-type, which most probably is to be associated with
the crushing of the chalk samples for the latter type of experi
ments, giving rise to new surfaces. In addition, slow equilibra
tion in the columns may play a role. The consistens between the
two different techniques, however, is unambiguously demonstrated
in the case of Co retardation on the heterogeneous columns.
Columns were made with crushed material from the core samples
4s-3 and 4s-6, the chalk being diluted by a factor 12.5 and 15,
respectively, by polystyrene spheres (cf. experimental section).
The column derived R« values (corrected tn pure consolidated -2 -2
chalk) were found to be 1.17*10 and 0.51*10 , respectively.
The corresponding values obtained in the batch-type experiments
were 1.72-10"2 and 1.02«10-2, respectively.
- 36 -
It should be noted ("able 6) that the trend observed, in the
batch-type experiments, indicating a slightly increased retar
dation by increasing depth below ground level is refound here.
4.5.2.1. Influence of bulk salt concentration in the eluent
From the data given in Table 6 it is noted that a slight depen
dence of the bulk salt concentration in the eluent apparently is
present. However, the here observed phenomena probably is only
•inor effects, whereas for More dramatic effects are observed
for salt concentrations below 0.5M. In fact, we were unable to
produce reasonable chromatograms eluting the columns with salt
solutions of concentrations below that level, as the signals be
came extreemly broad in addition to drastically decreased reten
tion factor. It is believed that the retention factors decreases
dramatically by further lowering of the salt concentration, in
good agreement with the recent investigations of Seitz et al.
(Seitz et al., 1979), who found extreemly small Rf(Cs) factors
in chalk formations. The same is most probably true for other
radionuclides. In fact pronounced tailing phenomena are observed
in the chromatograms even within the sodium chloride concentra
tion range studied here. In Fig. 20 the eluted Cs+, as func
tion of time and salt concentration from a 10 mm Erslev 4s-3
column is visualized.
4.5.2.2. Influence of complexing agents
Finally the effect of complexing agents shall briefly be men
tioned.
The possible influence of complexing agents was studied by in
jecting 50 ML samples of 134Cs*, 85Sr2*, 60Co2+, or 154Eu3+ sol
utions onto a homogeneous 10 mm Erslev 4s-3 column. Immediately
after, 50 ML of a Triplex (disodium salt of ethylenediamine tetra-
acetic acid) was injected and the retention times were then
studied. In the case of Triplex as complexing agent we found that
Co and 15 Eu in fact were eluted with the solvent front, 85 2+
whereas very minor retention of Sr is observed (R. s 0.95). 134 •
Mo influence on the Cs retention was observed. We conclude
- 37 -
134Cs*
Cs*/2M NaCt
B
mCs* 11
mCs*/4M NaCl
Pig. 20. Retention of Cs* In a 1 cm Erilev 4s-3 column as
function of the eluent sodium chloride concentration.
0.3 mL/min.).
(Flow rate:
on the basis of these experiments that in the presence of com-
plexing agents the retention of multi-valent radionuclides, such
as Co +, Sr +, and Eu decreases dramatically, whereas 134 +
mono-valent metal ions, as Cs , not are affected.
- 38 -
5. GEOCHEMCAL IMPLICATIONS
The geological formation overlying the top of the Erslev salt
dome consists mainly of chalk. The material is very fine grained
consisting mostly of skeletons of microorganisms from the cre
taceous period. The remains consist mainly of calcite more or
less enriched with magnesium (Kinard, 1980 and Chave, 1962); how
ever, in the present case the exact composition of the formation
has not been determined.
It has been reported that the presence of magnesium ions in the
aqueous phase inhibitates cementing processes in chalk formations
(Bjørlykke, 1977). Thus, the increase in magnesium concentration
with depth (Solgård and Skytte Jensen, 1981) has accordingly af
forded that the major part of the formation consists of rather
soft material of high porosity and consequently with a large free
surface area pr. unit volume.
5.1. Redox and pH conditions in the Erslev chalk formation
Since it has not been possible to collect uncontaminated samples
of ground water from the chalk formation to establish redox po
tentials and pH conditions, a model calculation has been carried
out in order to estimate the chemical conditions within the for
mation .
The analyses of pore water composition has been described in de
tail elsewhere (Solgård and Skytte Jensen, 1981). The chalk sam
ples were equilibrated for several hours with an artificial aque
ous phase before analysis of the liquid phase. Inspection of the
analytical data (Solgård and Skytte Jensen, 1981), revealed that
the measured concentrations of calcium seem to exceed calculated
equilibrium concentrations assuming simple dissolution of CaCO,.
Thus, in the present model it has been assumed that within the
chalk formation concentrations of different species are deter
mined by an equilibrium between calcium ions in solution, orig-
- 3» -
inating from dissolution of the salt and the cap-rock, the
latter consisting largely of anhydrite, i.e. CaSO., and calcite.
The presence of an increased calcium ion concentration in sol
ution influence the calcite solubility as well as equilibria in
volving carbonate ions. Other salts present in solution are con
sidered to be chemically inactive although contributing to the
overall ionic stregth. Through out the calculations an average
ionic strength equal to 1.0 has been assumed, applying available
equilibrium constants.
At an ionic strength equal to 1.0 the relevant equilibrium con
stants (given in Scheme 5-1) have been reported previously (Smith
and Kartell, 1976).
Scheme 5-1
[Ca2*]{C02~i
[CaOH*]/lCa2*l(0H"]
[CaHCO*]/tCa2*][HC03J
[H^HHCOp/CHjCOj)
[H*HC0 2 ~) / [HC0 3 )
[H*](0H~J
[CaC03]/[Ca2*][CoJ"l
10
10
10
10
10
10
10"
-8.01
0.(4
O.tl
-6.02
-».57
-13.7*
The principle of electroneutrality will in the present case lead
to Eq. 5.1
[H*J* 2[Ca2*]*[CaOH*MCaHCO*J •
2lCO2"j+[HCO3l*n[Xn"j*[0H"j (5.1)
where the concentration C* "J corresponds to the equivalents of
possible calcium salts ("C«Xn", e.g. X - CI, S04) dissolved in
the aqueous phase. In addition the stoechiometric equation (Eq.
5.2) is valid.
- 40 -
[Ca2+]+[CaOH+]+[CaHCO*J+[CaC03] =
[C03~l+[HCO3]+[H2C03]+[CaHCO*]+[CaC03]+n/2[Xn~] (5.2)
Combining the equations 5.1 and 5.2 the following expression
(Eq. 5.3) is obtained
[H+]-[CaOH+]+[CaHC03]+[HCO~]+2[H2C03]-[OH"] = 0 (5.3)
Substitution of the expressions given in Scheme 5-1 into Eq. 5.3
affords the following equation (Eq. 5.4)
[H+]-[Ca2+j3Kw/[H+]+YKs[H
+]/K2 +Ks[H*]/K2[Ca2+J+
2Kg[H+]2/K1K2[Ca
2+]-Kw/[H+] = 0 (5.4)
+ 2+
Eq. 5.4 can be solved for [H ] as function of [Ca ], which al
lows simultaneous calculation of concentrations of other species
present.
In Fig. 21 concentrations of the different species are depicted
as functions of total concentration of calcium in solution U.e.
cT • [Ca2+] + [CaOH*J + [CaHCOjl + [CaC03J). The total concen
tration, c„, calculated in this way corresponds closely to ana
lytical data for calcium concentrations in the aqueous phase
within the formation. These concentrations vary between 0.01 and
0.125M (Solgård and Skytte Jensen, 1981). Hence, it is accord
ingly possible to read the corresponding concentration ranges for
other species by application of Fig. 21. The chemical conditions
within the formation have been estimated as follows (Scheme 5-2).
According to these data the partial pressure of carbon dioxide
within the formation are estimated to be 10 atm.
As mentioned above the redox conditions within the formation has
not been measured, due to lack of autentic samples of ground
water. However, the observation (cf. footnote p. 5) that pyrite,
^.e. FeS-, is among the trace minerals found in the formation
- 41 -
Scheme 5 - 2
PH
t C 0 3 ~ ]
[HCO3]
[H 2 C0 3 ]
tCaCO,]
[CaOH ]
[CaHCO~]
RANGE
8 . 2 - 8 . 8
1 0 " 6 - i o - 7 - 2
I O ' 5 - i o - 5 - 8
= 1(T8
=1<T5
1<T5*8 - i o " 6 - 5
1 0 " 5 ' 8 - 1 0 " 6 * 5
-2 -
-4 -x
-6
-8
-10
CaSOj.
Ca2*
CO3" / " HCOJ C,„
coco3 ^ : : > ^ (
CaOH*..-::^--
_H2C03 • « • • * • * » • • • « «*«i
H* _L
-5 -4 -3 -2
log C(Ca) . 2 *
-1
CaOH , CaHCO,
CC-ij", HC(£, HJCOJ, and H* as function of total soluble Pig. 21. Equilibrium concentrations of Ca CaC03, CO3", HCO3, calcium in th« system calcit«/"CaXn". Equilibrium constants apply to th« condition of ionic strength equal to 1.0. Pull drawn lines correspond to the rang« found in th« cor« samples.
- 42 -
apparently fix the possible redox potential within rather narrow
limits close to the equilibrium line between S04 and HS /S in
a Eh-pH diagram (Fig. 22). The possible pH conditions calculated
(Scheme 5-2) lead to an expected redox potential at aproximately
-0.3 ± 0.1 volts, which identifies the area within the Eh-pH dia
grams most probably corresponding to the actual conditions to be
found in the Erslev chalk formation.
1.5
to
as
1 LI
ao
-as
-to 0 7 14
PH
Pig. 22. The hatched area indicate* the expected chemical condition* within the Erslev chalk formation.
5.2. Geochemistry of radionuclides
Examination of the stabil i ty diagrams previously presented in the report "The Geochemistry of Radionuclides with Long Halflifes" (Skytte Jensen, 1980), afford the spec^ation of the transuranium
i 1 i i i i i i i i—i—r
J I I I I I I I I I I l—L
- 43 -
elements as well as technetium under the above established con
ditions. The corresponding maximum solubilities of the single el
ements can be read from the diagrams displaying iso-concentration
curves within the Eh-pH field (Skytte Jensen, 1980).
The diagrams reveal that neptunium should be found in solution in
very minor concentrations only (in the order of 10~ M) apparent
ly as Np(0H)3 in equilibrium with the very slightly soluble NpO_.
Plutonium apparently appears in the oxidation state +3 as Pu(HO) -12
also in very low concentrations (in the order of 10 M) in equilibrium with the very slightly soluble PuO-.
Antericium will, as plutonium, exist in the oxidation state +3 as 2+ -4
Am(OH) , however, in concentrations up to 10 M in equilibrium
with the slightly soluble Am (OH).,. If, on the other hand, it can
be assured that the solubility product of Am,(CO,), is comparable -33 to that of Eu.(C0,)3, .i.e. 10 , the maximum solubility of
americium can be calculated to lie within the same range in the given system.
For technetium the tetravalent undissociated species TcO(OH)_
will be dominant under the actual conditions and will accordingly
exist in solution in very low concentrations only, in the order
of 10 M. The solid phase in equilibrium with the aqueous sol
ution apparently consists of the slightly soluble Tc02«
The retention experiments discussed above (cf. section 4.5) unam
biguously demonstrate that all cationic species, including Cs , 2+ 2* 3+
Sr , Co , and Eu more or less are adsorbed by the chalk,
whereas no retarding effect on anionic species like Cl~ and TcoT
are observed, suggesting ion-exchange as a possible mechanistic
approach, although simple precipitation reactions, as indicated
above, in addition may play an important role, especially in the
case of europium.
Ion exchange phenomena could a priori be ascribed to minor
amounts of clay material present in the chalk; on the other hand,
the relative insensitivity of the sorption on the bulk salt con-
- 44 -
contration, at least in the range above 0.5M, suggests that the
more specific surface binding mechanisms operate. It is known
that most mineral surfaces, including that of calcite, aquires a
negative charge at increased pH-values (Stumm and Morgan, 1970);
hence, they become capable of adsorbing or incorporating cationic
species into their surface layers. The available surface for ad
sorption of the present type of chalk may well approach several
square metres pr. gram and accordingly calcite itself may well be
the effective sorbing material responsible for the experimental
results given above. Supplementary column-type experiments to
134 +
study e.g. Cs retention on pure, precipitated CaCO- were car
ried out. The results obtained unequivocally confirm that calcium
carbonate most surely is the retarding material. In 1M sodium
chloride solution (as eluent) the following results were ob
tained: CaC03, Rf(Cs) = 0.23, Erslev 4s-3, Rf(Cs) = 0.26, Erslev
1s-5, Rf(Cs) = 0.23 (cf. Table 6).
Of the ions investigated, cesium, strontium, and technetium them
selves are important radionuclides in the nuclear waste. The tri
valent europium can be considered as a proper model compound for 3+ 3+
the trivalent actinides, in the present case Am and Pu , since
experiments as well as theory have shown similar chemical reac
tions for these ions. Additionally, rather close values for dif
ferent equilibrium constants have been estimated (Comprehensive
Inorganic Chemistry, 1973).
6. CONCLUSIONS
In the present study chalk samples from Erslev, Hors, Denmark
have been investigated predominantely by a liquid chromatographic
techique in order to establish permeability, porosity, disper
sion-, diffusion-, and sorption characteristics for the chalk
formation. In addition porosities have been estimated by density
measurements and sorption characteristics by batch type experi
ments .
- 45 -
It must strongly be emphasized that only 21 samples (6 from
Erslev Is, 3 from Erslev 2s, 6 from Erslev 3s, and 6 from Erslev
4s) have been available for these investigations, however, a rat
her clear-cut picture has been developed.
Although the samples in general exhibited high porosities, around
0.4, they were found to be rather impermeable, as the permeabili-—6 —8
ties were found in the range from 10 to 10 cm/sec. The permeabilities appear to decrease with increasing depth below ground level.
The flow-dispersion was, for experimental reasons, determined at
flow rates down to ca. 0.03 mL/min only, which apparently is very
much higher than the actual flow conditions within the formation. -5 2 The dispersion coefficients were determined to ca. 6»10 cm /sec,
close the diffusion coefficient in pure water. However, it is be
lieved that this value, due to the low permeability, may be up to
several orders of magnitude higher than the actual effective dif
fusion coefficient within the chalk.
22 • The apparent diffusion coefficients for Na in two representative chalk samples from the Erslev formation was determined. The actual magnitude was found to be up to ca. 10 times lower than the corresponding diffusion coefficient in pure water.
It was found that the chalk exihited a retarding effect on cat-+ 2+ 2+ 3+
ionic species such as Cs , Sr , Co , and Eu , whereas anionic
species as Cl" and TcO. were found to move with the water front,
i,.e. without retardation.
The likely geochemistry of radionuclides within the formation has
been discussed, leading to the conclusion that even in cases of
moderate water flows through the chalk formation, migration of
actinide cations, as Am and Pu effectively will be retarded.
Due to the extension of the Erslev chalk formation complementary
to its impermeability it seems probable that migration to the
surface may be dominated by diffusion.
- 46 -
ACKNOWLEDGEMENTS
This report has been worked out according to the agreement be
tween Ris# National Laboratory and ELSAM/ELKRAFT concerning ad
visory assistance frcai Ris# to ELSAM/ELKRAFT's management project,
Phase 2. The authors are indebted to Mrs. L. Halby and Mrs. G.
Larsen for valuable technical assistance.
REFERENCES
ATKINS, P.M. (1978). Physical Chemistry (Oxford University Press,
Oxford) Chapter 25.
BECK, R.E. and SCHULTZ, J.S. (1970). Science no, 1302-5.
BJØRLYKKE, K. (1977). Sedimentologi, Stratigrafi og Oliegeologi
(Universitetsforlaget, Bergen) 236 pp.
BO, P. and CARLSEN, L. (1981a). DIFM1G- A computer program for
calculation of diffusive migration of radionuclides through
multibarrier systems. Rise -M-2262. 33 pp.
BO, P. and CARLSEN, L. (1981b). European Appl. Res. Rept.-Nucl.
sci. Technol. 3, 813-73.
CHAVE, K.E. (1962). Limnol. Oceanog. 7, 218-23.
COMPREHENSIVE INORGANIC CHEMISTRY (1973). Vol. 5. Actinides. Ed.
by J.C. Bailar, Jr., H.J. Emeléus, Sir Ronald Nyholm and A.F.
Trotman-Dickenson (Perganom Press, Oxford).
DAHL, N.J. '1969). Grundvandbevcgelse (Teknisk Forlag, Copenhagen)
P. 29.
GRIM, R.E. (1968). clay Mineralogy (McGraw-Hill, New York) 220-222.
HARNED, H.S. and OWEN, B.B. (1963). The Physical ChemiBtry of
Electrolytic Solutions. 3. ed., 2. pr. (Reinhold, New York)
American Chemical Society. Monograph Series; 137. 803 pp.
KINARD, W.F. (1980). J. Chen. Ed. 5J7, 783-4.
MARTELL, A.E. and SMITH, R.M. (1976). Critical Stability Con
stants. Vol. 4. Inorganic Complexes (Plenum Press, New York).
- 47 -
MEARES, P. (1968). In: Diffusion in Polymers. Ed. by J. Crank and
G.S. Park (Academic Press, London) Chapter 10.
SEITZ, M.G., RICKERT, P.G., FRIED, S., FRIEDMAN, A.M. and
STEINDLER, M.J. (1979). Nucl. Technol. 44, 284-96.
SKYTTE JENSEN, B. (1980). The Geochemistry of Radionuclides with
long half-lives. Risø-R-430. 53 pp.
SOLGARD, P. and SKYTTE JENSEN, B. (1981). Chemical Analysis of
Pore Hater in Chalk Samples from Erslev, Mors, Denmark. Ris#
Chemistry Department. Contribution to the ELSAM/ELKRAFT man
agement project, Disposal of High-Level Waste from Nuclear
Power Plants in Denmark, Phase 2. Salt Dome Investigations.
Volume 2. Geology.
STUMM, W. and MORGAN, J.J. (1970). Aquatic Chemistry (Wiley, New
York) 585 pp.
YAU, W.W., KIRKLAND, J.J. and BLY, D.D. (1979). Modern Size -Ex
clusion Liquid Chromatography (Wiley, New York) P. 59.
Sales distributors: Jul. Gjellerup, Sølvgade 87, DK-1307 Copenhagen K, Denmark
Available on exchange from: * ?so Library, Ris* National Laboratory, ISBN 87-550-0785-* P. O. Box 49, DK-4000 Roskilde, Denmark ISSN 0106-2840